Responses

Create a response

OpenAI Responses API. Works with every chat model in the catalog, including Anthropic and Gemini ones: models whose upstream speaks the Responses protocol natively are served directly, the rest are translated to and from their own shape. The request you send and the response you get back are the same either way.

Server-side tools stay provider-specific — xAI's Agent Tools API uses this shape, so any Grok model accepts tools: [{"type": "web_search"}] / x_search / code_interpreter here.

Provider-native alternatives: POST /v1/messages for Anthropic (Native Formats / Anthropic) and POST /v1beta/models/{model}:generateContent for Gemini (Native Formats / Gemini).

post/responses

Request

  • Base URL
    https://api.orcarouter.ai/v1
  • URL
    https://api.orcarouter.ai/v1/responses
  • Auth
    HTTP bearer

Request body

modelstring required
instructionsstring
previous_response_idstring

ID of a prior response, to continue a multi-turn conversation without resending the full message history.

temperaturenumber
top_pnumber
max_output_tokensinteger
streamboolean
storeboolean

Whether the upstream may store the request/response. Allowed by default; channel setting disable_store can override.

metadataobject

Example request

{
  "model": "openai/gpt-4.1",
  "input": [
    {
      "content": [
        {
          "image_url": {
            "url": "https://example.com/photo.jpg"
          }
        }
      ]
    }
  ]
}

Response

OK

Changes

No recorded changes to this operation across all 2 revisions of this API.